Abstract:

Objective To establish a scientific method for recognition of the strokes sequence of Chinese character RU. Method Handwritings of the target character were collected and divided into training and test groups. The stroke sequences were reported by the writers. The pattern characteristics of the character were extracted from the training samples. The classifier was designed based on those characteristics， and further verified by the test samples. Results With the 352 valid samples obtained， 2 stroke sequences were reported. The pattern characteristics of the stroke movements were extracted and the discriminant function and rules were established. The recognition method was tested by the subjects. Conclusion A recognition method for sequencing of Chinese Character RU has been established， which has a negation rate of 11.4%， with an error rate of 4.6%.

Key words: questioned document examination, handwriting examination, strokes sequence, recognition

摘要：

目的 建立入字笔顺的科学识别方法。方法 收集实验笔迹样本分别作为训练样本和测试样本，由书写样本的被试者报告目标字的笔顺，依据训练样本提取入字的形态特征，在特征基础上建立分类器，使用测试样本评估分类器的性能。结果 共获得352份有效样本，被试者报告了2种入字笔顺，提取撇笔收笔动向等形态特征，拟定判别函数，建立判别规则，由测试者对识别方法进行测试。结论 建立了入字笔顺的识别方法，对于测试样本，拒识率为11.4%，测试错误率为4.6%。

关键词: 文件检验, 笔迹检验, 笔顺, 识别